A Piecewise Linear Network Classifier

Size: px
Start display at page:

Download "A Piecewise Linear Network Classifier"

Transcription

1 Proeedings of International Joint Conferene on eural etworks, Orlando, Florida, USA, August -7, 007 A Pieewise Linear etwork Classifier Abdul A. Abdurrab, Mihael T. Manry, Jiang Li, Sanjee S. Malalur and Robert G. Gore Abstrat A pieewise linear network is disussed whih lassifies -dimensional input etors. The network uses a distane measure to assign inoming input etors to an appropriate luster. Eah luster has a linear lassifier for generating lass disriminants. A training algorithm is desribed for generating the lusters and disriminants. Theorems are gien whih relate the network s performane to that of nearest neighbor and k-nearest neighbor lassifiers. It is shown that the error approahes Bayes error as the number of lusters and patterns per luster approah infinity. I. ITRODUCTIO eedforward neural nets with the uniersal approximation F property [,] mimi Bayes disriminants [3,4] and hae been suessfully used for many lassifiation tasks. Howeer, training time is slow, and onergene of the lassifiation error to Bayes error has not been shown. Suh onergene theorems do exist for nearest neighbor lassifiers (C) and k-nearest neighbor lassifiers (k-c) [5, 6], whih also hae the adantage of being easy to design. Howeer, the C and k-c are rarely used beause it is ery time-onsuming to apply them. Pieewise linear networks hae long been used for funtion approximation and lassifiation [7-] where speed of operation and simpliity are ery important. One approah is to train a multilayer pereptron (MLP) haing pieewise linear atiations [, 3]. This approah is useful in hardware implementations, and results in a ontinuous approximation. Howeer, if we are willing to gie up ontinuous approximation, a simpler pieewise linear network an be deised. In this paper we deelop a disontinuous pieewise linear network lassifier (PLC), based upon the work in [4]. The struture, operation and training of the PLC are desribed in setion II. In setion III the estimation of Bayesian a- posteriori probabilities by the PLC is disussed, along with its relation to the C and k-c. Simulations and performane omparisons are presented in setion IV. Manusript reeied January 3, 007 A. A. Abdurrab, M. T. Manry and S. S. Malalur are with the Department of Eletrial Engineering, Image Proessing and eural etworks Lab, Uniersity of Texas at Arlington, Arlington, TX USA ( abdulaziz.abdurrab@uta.edu, manry@uta.edu, smalalur@uta.edu). R. G. Gore is with Lokheed Martin, Mission Systems, Fort Worth, TX 760, USA. J. Li is with ECE Department and VMASC at Old Dominion Uniersity, orfolk VA 359, USA ( JLi@odu.edu). II. PIECEWISE LIEAR ETWORK CLASSIFIER A. Struture and Operation As shown in figure, the PLC has input elements in the first layer, the hidden units in the seond and outputs in the third. The -dimensional input etor x f has elements x f (n) with means and standard deiations µ(n) and σ(n) respetiely where n. First, x f is normalized as xf ( n) xf ( n) µ ( n) / σ( n) () The normalized (+)-dimensional etor x is formed by augmenting x f as x= ( x T :) T () f The hidden layer onsists of K lusters, eah haing an - dimensional mean etor m, where K. A Eulidean distane measure d( ) is used in lustering. Eah luster also has a weight matrix A of dimension by (+), where is the number of lasses in the lassifiation problem. d() Fig.. PLC Struture Gien an input etor x, we find luster index suh that d(x,m ) is minimized. Then we form the output etor y as y = A x (3) The estimate of the orret lass i is gien by i = arg max[ y ] (4) = = = K where y i is the i th element of output etor y and i. B. Training A lassifiation problem typially inoles a feature spae i i A A A K X/07/$ IEEE

2 with numerous feature etors or samples that hae to be assigned arious lass labels. In superised learning, the training dataset inludes the lass label, i, of eah of the feature etors. The label is transformed into an - dimensional target etor t suh that ( ) otherwise + b i = i p t pi = (5) b where p and b is a positie integer. Before the network is used for lassifiation, it has to be trained, whih inoles designing the weight matries A and the luster mean etors m. Clustering of the input etors x p is performed using the self-organizing map (SOM) [5]. Elaborating on (3), the elements of the output etor y, are alulated as + (6) y = a x pi in pn n= where a in denotes an element from the weight matrix A matrix belonging to row i and olumn n. The mean-squared error (MSE) for the th luster is then gien by ( ) + E = qi in qn ( ) t a x (7) i= n= where () is the number of feature etors in luster and t qi is the desired output whih ould be +b or -b. The error gradient with respet to a jm is gien by + = ajm On further simplifiation we get + = tqj xqm ajn xqnxqm ajm n= ( ) tqj ajnxqn x (8) qm n= (9) Equation (9) an be written in terms of auto-orrelation and ross-orrelation elements, r and respetiely, as + = ( j, m) ajnr( n, m) a (0) jm n= Equating the gradient in (0) to zero, in order to minimize the error, yields + (, ) (, ) j m = a r n m () n= The set of equations in () is soled using the Output Reset (OR) method [6-8] and a Shmidt based linear equation soler [9-0]. After the training is omplete, the PLC s m etors and A matries are fixed and an be used to lassify alidation data. jn III. EURAL ETWORKS AD THE BAYES OPTIMAL DISCRIMIAT FUCTIO A. eural networks and Bayes Disriminant Oer the past seeral years, neural networks hae been used for many tasks, inluding lassifiation and funtion approximation. There hae been many useful theoretial results onerning their apabilities inluding the following [4] theorem. Theorem: When neural net lassifiers are trained to minimize the mean-squared error (MSE), the MSE approahes a onstant alue plus the expeted squared error between the neural net output and Bayes disriminant, as the number of training patterns approahes infinity. Speifially, lim tpi yi ( p ) x = p= i= () i= ( i( x) i( x) ) E b y + a where a is a onstant, independent of p, t pi is defined in (5), and y i (x p ) is the i th output of the network. The Bayes disriminant b i (x), is the probability that the i th lass is orret, gien x, whih is written as P(i x). The aboe theorem, howeer, leaes room for some doubts: ) The neural network s probability of error is not bounded or shown to onerge to Bayes probability of error. ) The mean-square error in the theorem treats positie and negatie errors the same. Howeer, it is good if the orret lass disriminant is larger than desired, but bad if it is smaller [6-8]. This problem leads to suboptimal networks. 3) The theorem makes no use of the neural network s struture. It applies equally well to any disriminant designed by minimizing the MSE. B. earest eighbor and the PL Classifiers In the C, an input feature etor is assigned to the lass of its nearest training sample, usually determined ia the Eulidean distane. The feature spae is diided into onex regions or lusters eah haing a speifi lass label. Speifially, the distane measure performs a Voronoi tessellation of the -dimensional input feature spae. ow we onsider the relationship between the PLC and C. As K approahes infinity, the onex Voronoi ells in the feature spae get smaller in olume and the optimal deision boundaries in eah luster beome linear. Hene, eah luster an hae its own linear disriminant and oerall, a more omplex deision boundary is ahieed. Therefore, for a gien alue of K, the PLC should perform better than the C. We begin to address this idea in the following lemmas. Lemma : If a PLC and C hae the same distane measure and idential luster mean etors, the PLC has at

3 least as good a performane as the C. Proof: Sine the PLC s augmented input etor x inludes the onstant ; the PLC s output etor y an hae a for the same lass as piked by the C. Other lass outputs an be 0. Suppose that the th C luster belongs to lass i. If we would like the PLC s th luster to always map patterns to lass i, the elements of the by (+) matrix A are defined as m= i and n= + a ( m, n) = (3) 0 otherwise When a PLC s A matrix initially performs worse than the C for a gien th luster, replae A with A as defined in (3). Lemma : If a PLC and a C hae the same distane measure and idential luster mean etors, then as the number of lusters, K and patterns, approah infinity, P P P P (4) e( PLC) e( C) where P e(plc), P e(c) and P respetiely denote the PLC, C and Bayes probabilities of error. Proof: As K approahes infinity, we know that [5, 6] P P P (5) ( ) e C Using lemma and (5) yields the result of (4). C. Conergene of the PLC Error Probability We now wish to be more speifi and ealuate the aerage probability of error for the PLC, as the amount of training data inreases. First, reall the following result desribing the onergene of the k-c probability of error, P e(k-c). Here k is the number of nearest training etors or samples to the test sample x. Lemma 3 [5, 6]: As k and ( /k) approah infinity, the k- C an be iewed as an attempt to estimate the a- posteriori probabilities from the training samples. Under this ondition, k-c beomes optimal and lim P = P (6) k, k e(k-c) We now inestigate the relationship between Bayes disriminant and the PLC in more detail. We start by obsering that the lassifiation error for the th luster of a PLC is gien by (5). Let us take into onsideration only the feature etors that hae been assigned lass label i. In this ase, the lassifiation error would be gien by + Ei, = tqi ainxqn (7) n= The partial deriatie of the error with respet to the weight matrix elements, a im, is gien by: + = tqi ainxqn xqm aim () (8) n= Before going any further, we onsider how letting K and () approah infinity effets the range and ariability of the input feature etors x. Assuming the feature spae is a bounded ompat subset of R, eah luster falls within a ell of the Voronoi tessellation. As K and () simultaneously inrease towards infinity, the maximum radius of eah luster dereases towards zero. The range of alues the elements of a feature etor in a luster an take dereases so muh that they beome onstant, i.e, ariability within a luster approahes zero. Within a gien luster, as x beomes independent of the pattern number p, we an replae x pn with mean etor element m n in (8). We also equate the gradient to zero, so as to minimize the MSE. Equation (8) redues to + tqi ainmn mm 0 () = (9) n= whih beomes + tqi ainmn = 0 (0) n= Sine the terms a in and m n are onstants, we an replae the sum oer n by the single onstant a i(+) whih yields This an be written as whih yields tqi a i( + ) = 0 () a tqi = () a () i( + ) qi ( ) = i + t (3) Without loss of generality, we an represent t qi as t = δ i q i (4) qi ( ( ) ) where δ is the Kroneker delta funtion. In turn (3) beomes a ( + ) i = δ ( i ( q) i) (5) δ(i (q) i) equals if i (q) equals i, so, the numerator summation equals the number of training etors in luster that belong to lass i. This number ould be represented by (,i), hene reduing (5) to (,) i ai( ) = + (6) whih onerges to the a-posteriori probability P(i x). For a gien luster, a i(+) is largest for the lass i whih has the most input etors in the luster. This is preisely how the k-c makes deisions. So, eah PLC luster with () members emulates a k-c deision with k = (). Summarizing the results aboe, we hae the following

4 lemma. Lemma 4: As K and () approah infinity, the output of a PLC approximates the a-posteriori probability funtions of the lass labels, gien the input etor. Under this ondition, the PLC hene beomes optimal and lim P e(plc) = P (7) K, ( ) window. For this omparison too, the PLC and C are allowed to design and use their own set of luster enter etors. The results of this omparison are shown in fig. 4. As in the preious examples, the PLC outperforms the C. IV. SIMULATIO RESULTS AD COMPARISOS In this setion, we ompare the PLC with the C, MLP and a pieewise linear MLP (PLMLP) [] using different data sets. Eah data set onsists of a training file and a alidation file. A. PLC s C Here, we ompare the PLC to a C haing the same number of luster mean etors. Three data sets are used for the omparison. ) grng [] This geometri shape reognition data set has four shapes haing different degrees of rotation, saling, translation, and oblique distortions. The shapes are ellipse, triangle, quadrilateral, and pentagon. Eah image is 64 x 64 pixels. The input etors ontain 6 ring-wedge features. For this omparison, the C and PLC, both use the same set of luster enter etors for lassifiation. In fig., the C performs poorly beause it is fored to use the same lusters as the PLC, for this example only. ) gong [] The raw data onsists of images from hand printed numerals olleted from 3,000 people by the Internal Reenue Serie. Images are 3 by 4 binary matries, whih are saled to remoe harater size ariation. The feature set ontains 6 elements. The 0 lasses orrespond to the Arabi numerals. For this omparison, the PLC and C are allowed to design and use their own sets of luster enter etors. The results of this omparison are shown in fig. 3. The linear lassifier in eah PLC luster allows the PLC to outperform the C. 3) omf8 [3] This data set onsists of texture features orresponding to an image segmentation problem. Eah segmented region is separately histogram equalized to 0 leels. Then the joint probability density of pairs of pixels separated by a gien distane and a gien diretion is estimated. We use 0, 90, 80, 70 degrees for the diretions and, 3, and 5 pixels for the separations. The density estimates are omputed for eah lassifiation window. For eah separation, the oourrenes for the four diretions are folded together to form a triangular matrix. From eah of the resulting three matries, six features are omputed: angular seond moment, ontrast, entropy, orrelation, and the sums of the main diagonal and the first off diagonal. This results in 8 features for eah lassifiation Fig.. Comparison of PLC and C for shape reognition data Fig. 3. Comparison of PLC and C for numeral reognition problem B. PLC s MLP s PLMLP Here, we ompare the PLC to a sigmoidal MLP and to a PLMLP []. Unlike the PLC, the PLMLP produes ontinuous mappings. Two data sets are used for the omparison. ) F7C This prognostis data set onsists of parameters that are aailable in the basi health usage monitoring system (HUMS) by Bell Heliopter Textron. The data was obtained from the M430 flight load leel surey onduted in Mirabel Canada. Eah input etor ontains 7 elements. The 39 lasses represent different

5 flight maneuers like taking off, landing, turning right or left et. The results of this omparison are shown in fig. 5. In this example, the PLC outperforms both the MLP and PLMLP. Fig. 6. Comparison of PLC, PLMLP and MLP for phoneme reognition problem Fig. 4. Comparison of PLC and C for segmentation problem C. Comparisons Based On the umbers of Multiplies Here, we re-do the plots of subsetion B, so that the three full trained networks are ompared based upon the numbers of multiplies required to proess one input etor. For the sigmoidal MLP it is assumed that two multiplies are required to ealuate the atiation. The PLMLP, in ontrast, requires only one multiply for alulating the atiation. Figures 7 and 8 show that a PLC requires fewer multipliations to obtain the same lassifiation error as ompared to the MLP and PLMLP. Fig. 5. Comparison of PLC, PLMLP and MLP for prognostis ) speeh - Speeh samples are first pre-emphasized and are onerted into frequeny domain by taking their DFT. They are then passed through Mel filter banks and the inerse DFT is applied on the output to get Mel- Frequeny Cepstrum Coeffiients (MFCC). Eah of MFCC(n), MFCC(n)-MFCC(n-) and MFCC(n)- MFCC(n-) would hae 3 features, whih results in a total of 39 features. Eah lass orresponds to a phoneme. The results of this omparison are shown in fig. 6. The PLC outperforms the PLMLP and the MLP for larger networks. V. COCLUSIOS In this paper we hae desribed a pieewise linear network lassifier (PLC) and gien a training algorithm for it. This algorithm has fast onergene speed beause sets of linear equations are soled in eah luster. Using existing theorems for the C and k-c, we hae proed that the output of the PLC approximates a Bayes optimal disriminant when trained to minimize the mean square error (MSE). Using seeral data sets, we hae shown that the PLC often performs better than equialent Cs, MLPs, and PLMLPs. It an also be stated that the PLC always trains faster than the MLP and often faster than the C, whih usually requires a lot more lusters. Future work will ompare arious alternatie methods for training the linear lassifiers. PL pruning tehniques [4] will also be extended to the PLC.

6 Fig. 7. Comparison of PLC, PLMLP and MLP for prognostis data, based on number of multiplies Fig. 8. Comparison of PLC, PLMLP and MLP for phoneme reognition problem, based on number of multiplies REFERECES [] K. Hornik, M. Stinhombe, and H. White, Multilayer Feedforward etworks Are Uniersal Approximators, eural etworks, Vol., o. 5, 989, pp [] K. Hornik, M. Stinhombe, and H. White, Uniersal Approximation of an Unknown Mapping and its Deriaties Using Multilayer Feedforward etworks, eural etworks, ol. 3, 990, pp [3] Mihael D. Rihard and Rihard P. Lippman, eural etwork Classifiers estimate Bayesian a-posteriori probabilities, eural Computation, ol. 3, no. 4, pp , 99. [4] Dennis W. Ruk, Steen K. Rogers, Matthew Kabrisky, Mark E. Oxley and Brue W. Suter, The Multilayer Pereptron as an approximation to a Bayes optimal disriminant funtion, IEEE Trans eural etworks, T-(4):96-98, 990. [5] K. Fukunaga, Introdution to Statistial Pattern Reognition, nd ed., Aademi Press, 990. [6] R.O. Duda, P.E. Hart, and D.G. Stork, Pattern Classifiation, John Wiley & Sons, nd ed., 00. [7] L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stone, Classifiation and Regression Trees, Wadsworth, Belmont, CA, 984. [8] D. B. Fogel, An information riterion for optimal neural network seletion, IEEE Trans. eural etworks, ol., no. 5, pp , Sept.99. [9] J. H. Friedman, Multiariate adaptie regression splines, Annals of Statistis, ol. 9, no., pp. -4, 99. [0] S. Subbarayan, K. Kim, M. T. Manry, V. Dearajan and H. Chen, Modular neural network arhiteture using pieewise linear mapping, 30th Asilomar Conferene on Signals, Systems & Computers, ol., pp. 7-75, o [] W. Li, J.-. Lin, and R. Unbehauen, Canonial representation of pieewise polynomial funtions with nondegenerate linear domain partitions, IEEE Trans. Ciruits and Systems I: Fundamental Theory and Appliations, ol. 45, no. 8, pp , Aug [] D.R. Hush and B. Horne, Effiient algorithms for funtion approximation with pieewise linear sigmoidal networks, IEEE Trans. eural etworks, Vol. 9, o. 6, pp. 9-4, 998. [3] E.F. Gad, A.F. Atiya, S. Shaheen, A. El-Dessouki, A new algorithm for learning in pieeswise-linear neural networks, eural etworks 3 (000), pp [4] H. Chandrasekaran, J. Li, W.H. Delashmit, P.L. arasimha, C. Yu and M.T. Manry, Conergent Design of Pieewise Linear eural etworks, eurocomputing, ol. 70, 007, pp [5] T. Kohonen, Self-Organization and Assoiatie Memory, nd ed., Springer-Verlag, 987. [6] L-M Liu, M.T. Manry, F. Amar, M.S. Dawson, and A.K. Fung, "Iteratie Improement of Image Classifiers Using Relaxation," Conferene Reord of the Twenty-Eighth Annual Asilomar Conferene on Signals, Systems, and Computers, ol., 0/3/94 to //94, pp [7] Jiang Li, Mihael T. Manry, Li-Min Liu, Changhua Yu, and John Wei, Iteratie Improement of eural Classifiers, Proeedings of the Seenteenth International Conferene of the Florida AI Researh Soiety, May 004, pp [8] R.G. Gore, Jiang Li, Mihael T. Manry, Li-Min Liu, Changhua Yu, and John Wei, "Iteratie Design of eural etwork Classifiers through Regression". International Journal on Artifiial Intelligene Tools, Vol. 4, os. & (005) pp [9] F. J. Maldonado, M. T. Manry, Tae-Hoon Kim, "Finding optimal neural network basis funtion subsets using the Shmidt proedure", Proeedings of the International Joint Conferene on eural etworks, ol., pp , July 003. [0] F. J. Maldonado, M. T. Manry, "Optimal Pruning of Feed-forward eural etworks Based upon the Shmidt Proedure," The 36th Asilomar Conferene on Signals, Systems, & Computers '0, pp [] H. C. Yau, M. T. Manry, "Iteratie Improement of a earest eighbor Classifier," eural etworks, Vol. 4, 99, pp [] W. Gong, H. C. Yau, and M. T. Manry, "on-gaussian Feature Analyses Using a eural etwork," Progress in eural etworks, ol., 994, pp [3] R.R. Bailey, E. J. Pettit, R. T. Borohoff, M. T. Manry, and X. Jiang, "Automati Reognition of USGS Land Use/Coer Categories Using Statistial and eural etwork Classifiers," Proeedings of SPIE OE/Aerospae and Remote Sensing, April -6, 993, Orlando Florida.

A Novel Validity Index for Determination of the Optimal Number of Clusters

A Novel Validity Index for Determination of the Optimal Number of Clusters IEICE TRANS. INF. & SYST., VOL.E84 D, NO.2 FEBRUARY 2001 281 LETTER A Novel Validity Index for Determination of the Optimal Number of Clusters Do-Jong KIM, Yong-Woon PARK, and Dong-Jo PARK, Nonmembers

More information

An Optimized Approach on Applying Genetic Algorithm to Adaptive Cluster Validity Index

An Optimized Approach on Applying Genetic Algorithm to Adaptive Cluster Validity Index IJCSES International Journal of Computer Sienes and Engineering Systems, ol., No.4, Otober 2007 CSES International 2007 ISSN 0973-4406 253 An Optimized Approah on Applying Geneti Algorithm to Adaptive

More information

Abstract. Key Words: Image Filters, Fuzzy Filters, Order Statistics Filters, Rank Ordered Mean Filters, Channel Noise. 1.

Abstract. Key Words: Image Filters, Fuzzy Filters, Order Statistics Filters, Rank Ordered Mean Filters, Channel Noise. 1. Fuzzy Weighted Rank Ordered Mean (FWROM) Filters for Mixed Noise Suppression from Images S. Meher, G. Panda, B. Majhi 3, M.R. Meher 4,,4 Department of Eletronis and I.E., National Institute of Tehnology,

More information

A {k, n}-secret Sharing Scheme for Color Images

A {k, n}-secret Sharing Scheme for Color Images A {k, n}-seret Sharing Sheme for Color Images Rastislav Luka, Konstantinos N. Plataniotis, and Anastasios N. Venetsanopoulos The Edward S. Rogers Sr. Dept. of Eletrial and Computer Engineering, University

More information

The Minimum Redundancy Maximum Relevance Approach to Building Sparse Support Vector Machines

The Minimum Redundancy Maximum Relevance Approach to Building Sparse Support Vector Machines The Minimum Redundany Maximum Relevane Approah to Building Sparse Support Vetor Mahines Xiaoxing Yang, Ke Tang, and Xin Yao, Nature Inspired Computation and Appliations Laboratory (NICAL), Shool of Computer

More information

NONLINEAR BACK PROJECTION FOR TOMOGRAPHIC IMAGE RECONSTRUCTION. Ken Sauer and Charles A. Bouman

NONLINEAR BACK PROJECTION FOR TOMOGRAPHIC IMAGE RECONSTRUCTION. Ken Sauer and Charles A. Bouman NONLINEAR BACK PROJECTION FOR TOMOGRAPHIC IMAGE RECONSTRUCTION Ken Sauer and Charles A. Bouman Department of Eletrial Engineering, University of Notre Dame Notre Dame, IN 46556, (219) 631-6999 Shool of

More information

Cluster-Based Cumulative Ensembles

Cluster-Based Cumulative Ensembles Cluster-Based Cumulative Ensembles Hanan G. Ayad and Mohamed S. Kamel Pattern Analysis and Mahine Intelligene Lab, Eletrial and Computer Engineering, University of Waterloo, Waterloo, Ontario N2L 3G1,

More information

the data. Structured Principal Component Analysis (SPCA)

the data. Structured Principal Component Analysis (SPCA) Strutured Prinipal Component Analysis Kristin M. Branson and Sameer Agarwal Department of Computer Siene and Engineering University of California, San Diego La Jolla, CA 9193-114 Abstrat Many tasks involving

More information

Smooth Trajectory Planning Along Bezier Curve for Mobile Robots with Velocity Constraints

Smooth Trajectory Planning Along Bezier Curve for Mobile Robots with Velocity Constraints Smooth Trajetory Planning Along Bezier Curve for Mobile Robots with Veloity Constraints Gil Jin Yang and Byoung Wook Choi Department of Eletrial and Information Engineering Seoul National University of

More information

Boosted Random Forest

Boosted Random Forest Boosted Random Forest Yohei Mishina, Masamitsu suhiya and Hironobu Fujiyoshi Department of Computer Siene, Chubu University, 1200 Matsumoto-ho, Kasugai, Aihi, Japan {mishi, mtdoll}@vision.s.hubu.a.jp,

More information

Pipelined Multipliers for Reconfigurable Hardware

Pipelined Multipliers for Reconfigurable Hardware Pipelined Multipliers for Reonfigurable Hardware Mithell J. Myjak and José G. Delgado-Frias Shool of Eletrial Engineering and Computer Siene, Washington State University Pullman, WA 99164-2752 USA {mmyjak,

More information

A Coarse-to-Fine Classification Scheme for Facial Expression Recognition

A Coarse-to-Fine Classification Scheme for Facial Expression Recognition A Coarse-to-Fine Classifiation Sheme for Faial Expression Reognition Xiaoyi Feng 1,, Abdenour Hadid 1 and Matti Pietikäinen 1 1 Mahine Vision Group Infoteh Oulu and Dept. of Eletrial and Information Engineering

More information

Learning Convention Propagation in BeerAdvocate Reviews from a etwork Perspective. Abstract

Learning Convention Propagation in BeerAdvocate Reviews from a etwork Perspective. Abstract CS 9 Projet Final Report: Learning Convention Propagation in BeerAdvoate Reviews from a etwork Perspetive Abstrat We look at the way onventions propagate between reviews on the BeerAdvoate dataset, and

More information

A Novel Bit Level Time Series Representation with Implication of Similarity Search and Clustering

A Novel Bit Level Time Series Representation with Implication of Similarity Search and Clustering A Novel Bit Level Time Series Representation with Impliation of Similarity Searh and lustering hotirat Ratanamahatana, Eamonn Keogh, Anthony J. Bagnall 2, and Stefano Lonardi Dept. of omputer Siene & Engineering,

More information

arxiv: v1 [cs.db] 13 Sep 2017

arxiv: v1 [cs.db] 13 Sep 2017 An effiient lustering algorithm from the measure of loal Gaussian distribution Yuan-Yen Tai (Dated: May 27, 2018) In this paper, I will introdue a fast and novel lustering algorithm based on Gaussian distribution

More information

Gray Codes for Reflectable Languages

Gray Codes for Reflectable Languages Gray Codes for Refletable Languages Yue Li Joe Sawada Marh 8, 2008 Abstrat We lassify a type of language alled a refletable language. We then develop a generi algorithm that an be used to list all strings

More information

A DYNAMIC ACCESS CONTROL WITH BINARY KEY-PAIR

A DYNAMIC ACCESS CONTROL WITH BINARY KEY-PAIR Malaysian Journal of Computer Siene, Vol 10 No 1, June 1997, pp 36-41 A DYNAMIC ACCESS CONTROL WITH BINARY KEY-PAIR Md Rafiqul Islam, Harihodin Selamat and Mohd Noor Md Sap Faulty of Computer Siene and

More information

The Alpha Torque and Quantum Physics

The Alpha Torque and Quantum Physics The Alpha Torque and Quantum Physis Zhiliang Cao, Henry Cao williamao15000@yahoo.om, henry.gu.ao@gmail.om July 18, 010 Abstrat In the enter of the unierse, there isn t a super massie blak hole or any speifi

More information

KERNEL SPARSE REPRESENTATION WITH LOCAL PATTERNS FOR FACE RECOGNITION

KERNEL SPARSE REPRESENTATION WITH LOCAL PATTERNS FOR FACE RECOGNITION KERNEL SPARSE REPRESENTATION WITH LOCAL PATTERNS FOR FACE RECOGNITION Cuiui Kang 1, Shengai Liao, Shiming Xiang 1, Chunhong Pan 1 1 National Laboratory of Pattern Reognition, Institute of Automation, Chinese

More information

HEXA: Compact Data Structures for Faster Packet Processing

HEXA: Compact Data Structures for Faster Packet Processing Washington University in St. Louis Washington University Open Sholarship All Computer Siene and Engineering Researh Computer Siene and Engineering Report Number: 27-26 27 HEXA: Compat Data Strutures for

More information

Extracting Partition Statistics from Semistructured Data

Extracting Partition Statistics from Semistructured Data Extrating Partition Statistis from Semistrutured Data John N. Wilson Rihard Gourlay Robert Japp Mathias Neumüller Department of Computer and Information Sienes University of Strathlyde, Glasgow, UK {jnw,rsg,rpj,mathias}@is.strath.a.uk

More information

Transition Detection Using Hilbert Transform and Texture Features

Transition Detection Using Hilbert Transform and Texture Features Amerian Journal of Signal Proessing 1, (): 35-4 DOI: 1.593/.asp.1.6 Transition Detetion Using Hilbert Transform and Texture Features G. G. Lashmi Priya *, S. Domni Department of Computer Appliations, National

More information

Weak Dependence on Initialization in Mixture of Linear Regressions

Weak Dependence on Initialization in Mixture of Linear Regressions Proeedings of the International MultiConferene of Engineers and Computer Sientists 8 Vol I IMECS 8, Marh -6, 8, Hong Kong Weak Dependene on Initialization in Mixture of Linear Regressions Ryohei Nakano

More information

Performance of Histogram-Based Skin Colour Segmentation for Arms Detection in Human Motion Analysis Application

Performance of Histogram-Based Skin Colour Segmentation for Arms Detection in Human Motion Analysis Application World Aademy of Siene, Engineering and Tehnology 8 009 Performane of Histogram-Based Skin Colour Segmentation for Arms Detetion in Human Motion Analysis Appliation Rosalyn R. Porle, Ali Chekima, Farrah

More information

1. Introduction. 2. The Probable Stope Algorithm

1. Introduction. 2. The Probable Stope Algorithm 1. Introdution Optimization in underground mine design has reeived less attention than that in open pit mines. This is mostly due to the diversity o underground mining methods and omplexity o underground

More information

Calculation of typical running time of a branch-and-bound algorithm for the vertex-cover problem

Calculation of typical running time of a branch-and-bound algorithm for the vertex-cover problem Calulation of typial running time of a branh-and-bound algorithm for the vertex-over problem Joni Pajarinen, Joni.Pajarinen@iki.fi Otober 21, 2007 1 Introdution The vertex-over problem is one of a olletion

More information

Rotation Invariant Spherical Harmonic Representation of 3D Shape Descriptors

Rotation Invariant Spherical Harmonic Representation of 3D Shape Descriptors Eurographis Symposium on Geometry Proessing (003) L. Kobbelt, P. Shröder, H. Hoppe (Editors) Rotation Invariant Spherial Harmoni Representation of 3D Shape Desriptors Mihael Kazhdan, Thomas Funkhouser,

More information

A Load-Balanced Clustering Protocol for Hierarchical Wireless Sensor Networks

A Load-Balanced Clustering Protocol for Hierarchical Wireless Sensor Networks International Journal of Advanes in Computer Networks and Its Seurity IJCNS A Load-Balaned Clustering Protool for Hierarhial Wireless Sensor Networks Mehdi Tarhani, Yousef S. Kavian, Saman Siavoshi, Ali

More information

2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media,

2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any urrent or future media, inluding reprinting/republishing this material for advertising

More information

Detection and Recognition of Non-Occluded Objects using Signature Map

Detection and Recognition of Non-Occluded Objects using Signature Map 6th WSEAS International Conferene on CIRCUITS, SYSTEMS, ELECTRONICS,CONTROL & SIGNAL PROCESSING, Cairo, Egypt, De 9-31, 007 65 Detetion and Reognition of Non-Oluded Objets using Signature Map Sangbum Park,

More information

Accommodations of QoS DiffServ Over IP and MPLS Networks

Accommodations of QoS DiffServ Over IP and MPLS Networks Aommodations of QoS DiffServ Over IP and MPLS Networks Abdullah AlWehaibi, Anjali Agarwal, Mihael Kadoh and Ahmed ElHakeem Department of Eletrial and Computer Department de Genie Eletrique Engineering

More information

Discrete sequential models and CRFs. 1 Case Study: Supervised Part-of-Speech Tagging

Discrete sequential models and CRFs. 1 Case Study: Supervised Part-of-Speech Tagging 0-708: Probabilisti Graphial Models 0-708, Spring 204 Disrete sequential models and CRFs Leturer: Eri P. Xing Sribes: Pankesh Bamotra, Xuanhong Li Case Study: Supervised Part-of-Speeh Tagging The supervised

More information

TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 1. Preserving Privacy in Distributed Systems

TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 1. Preserving Privacy in Distributed Systems TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS Presering Priay in Distributed Systems Yuriy Brun and Nenad Medidoi, Member, IEEE Computer Soiety Abstrat We present stile, a tehnique for distributing

More information

INTERPOLATED AND WARPED 2-D DIGITAL WAVEGUIDE MESH ALGORITHMS

INTERPOLATED AND WARPED 2-D DIGITAL WAVEGUIDE MESH ALGORITHMS Proeedings of the COST G-6 Conferene on Digital Audio Effets (DAFX-), Verona, Italy, Deember 7-9, INTERPOLATED AND WARPED -D DIGITAL WAVEGUIDE MESH ALGORITHMS Vesa Välimäki Lab. of Aoustis and Audio Signal

More information

Graph-Based vs Depth-Based Data Representation for Multiview Images

Graph-Based vs Depth-Based Data Representation for Multiview Images Graph-Based vs Depth-Based Data Representation for Multiview Images Thomas Maugey, Antonio Ortega, Pasal Frossard Signal Proessing Laboratory (LTS), Eole Polytehnique Fédérale de Lausanne (EPFL) Email:

More information

Gradient based progressive probabilistic Hough transform

Gradient based progressive probabilistic Hough transform Gradient based progressive probabilisti Hough transform C.Galambos, J.Kittler and J.Matas Abstrat: The authors look at the benefits of exploiting gradient information to enhane the progressive probabilisti

More information

A scheme for racquet sports video analysis with the combination of audio-visual information

A scheme for racquet sports video analysis with the combination of audio-visual information A sheme for raquet sports video analysis with the ombination of audio-visual information Liyuan Xing a*, Qixiang Ye b, Weigang Zhang, Qingming Huang a and Hua Yu a a Graduate Shool of the Chinese Aadamy

More information

A New RBFNDDA-KNN Network and Its Application to Medical Pattern Classification

A New RBFNDDA-KNN Network and Its Application to Medical Pattern Classification A New RBFNDDA-KNN Network and Its Appliation to Medial Pattern Classifiation Shing Chiang Tan 1*, Chee Peng Lim 2, Robert F. Harrison 3, R. Lee Kennedy 4 1 Faulty of Information Siene and Tehnology, Multimedia

More information

Capturing Large Intra-class Variations of Biometric Data by Template Co-updating

Capturing Large Intra-class Variations of Biometric Data by Template Co-updating Capturing Large Intra-lass Variations of Biometri Data by Template Co-updating Ajita Rattani University of Cagliari Piazza d'armi, Cagliari, Italy ajita.rattani@diee.unia.it Gian Lua Marialis University

More information

Drawing lines. Naïve line drawing algorithm. drawpixel(x, round(y)); double dy = y1 - y0; double dx = x1 - x0; double m = dy / dx; double y = y0;

Drawing lines. Naïve line drawing algorithm. drawpixel(x, round(y)); double dy = y1 - y0; double dx = x1 - x0; double m = dy / dx; double y = y0; Naïve line drawing algorithm // Connet to grid points(x0,y0) and // (x1,y1) by a line. void drawline(int x0, int y0, int x1, int y1) { int x; double dy = y1 - y0; double dx = x1 - x0; double m = dy / dx;

More information

Supplementary Material: Geometric Calibration of Micro-Lens-Based Light-Field Cameras using Line Features

Supplementary Material: Geometric Calibration of Micro-Lens-Based Light-Field Cameras using Line Features Supplementary Material: Geometri Calibration of Miro-Lens-Based Light-Field Cameras using Line Features Yunsu Bok, Hae-Gon Jeon and In So Kweon KAIST, Korea As the supplementary material, we provide detailed

More information

Algorithms for External Memory Lecture 6 Graph Algorithms - Weighted List Ranking

Algorithms for External Memory Lecture 6 Graph Algorithms - Weighted List Ranking Algorithms for External Memory Leture 6 Graph Algorithms - Weighted List Ranking Leturer: Nodari Sithinava Sribe: Andi Hellmund, Simon Ohsenreither 1 Introdution & Motivation After talking about I/O-effiient

More information

Naïve Bayesian Rough Sets Under Fuzziness

Naïve Bayesian Rough Sets Under Fuzziness IJMSA: Vol. 6, No. 1-2, January-June 2012, pp. 19 25 Serials Publiations ISSN: 0973-6786 Naïve ayesian Rough Sets Under Fuzziness G. GANSAN 1,. KRISHNAVNI 2 T. HYMAVATHI 3 1,2,3 Department of Mathematis,

More information

Outline: Software Design

Outline: Software Design Outline: Software Design. Goals History of software design ideas Design priniples Design methods Life belt or leg iron? (Budgen) Copyright Nany Leveson, Sept. 1999 A Little History... At first, struggling

More information

Approximate logic synthesis for error tolerant applications

Approximate logic synthesis for error tolerant applications Approximate logi synthesis for error tolerant appliations Doohul Shin and Sandeep K. Gupta Eletrial Engineering Department, University of Southern California, Los Angeles, CA 989 {doohuls, sandeep}@us.edu

More information

Cluster Centric Fuzzy Modeling

Cluster Centric Fuzzy Modeling 10.1109/TFUZZ.014.300134, IEEE Transations on Fuzzy Systems TFS-013-0379.R1 1 Cluster Centri Fuzzy Modeling Witold Pedryz, Fellow, IEEE, and Hesam Izakian, Student Member, IEEE Abstrat In this study, we

More information

An Efficient and Scalable Approach to CNN Queries in a Road Network

An Efficient and Scalable Approach to CNN Queries in a Road Network An Effiient and Salable Approah to CNN Queries in a Road Network Hyung-Ju Cho Chin-Wan Chung Dept. of Eletrial Engineering & Computer Siene Korea Advaned Institute of Siene and Tehnology 373- Kusong-dong,

More information

Automatic Physical Design Tuning: Workload as a Sequence Sanjay Agrawal Microsoft Research One Microsoft Way Redmond, WA, USA +1-(425)

Automatic Physical Design Tuning: Workload as a Sequence Sanjay Agrawal Microsoft Research One Microsoft Way Redmond, WA, USA +1-(425) Automati Physial Design Tuning: Workload as a Sequene Sanjay Agrawal Mirosoft Researh One Mirosoft Way Redmond, WA, USA +1-(425) 75-357 sagrawal@mirosoft.om Eri Chu * Computer Sienes Department University

More information

Implementing Load-Balanced Switches With Fat-Tree Networks

Implementing Load-Balanced Switches With Fat-Tree Networks Implementing Load-Balaned Swithes With Fat-Tree Networks Hung-Shih Chueh, Ching-Min Lien, Cheng-Shang Chang, Jay Cheng, and Duan-Shin Lee Department of Eletrial Engineering & Institute of Communiations

More information

Relevance for Computer Vision

Relevance for Computer Vision The Geometry of ROC Spae: Understanding Mahine Learning Metris through ROC Isometris, by Peter A. Flah International Conferene on Mahine Learning (ICML-23) http://www.s.bris.a.uk/publiations/papers/74.pdf

More information

A Dictionary based Efficient Text Compression Technique using Replacement Strategy

A Dictionary based Efficient Text Compression Technique using Replacement Strategy A based Effiient Text Compression Tehnique using Replaement Strategy Debashis Chakraborty Assistant Professor, Department of CSE, St. Thomas College of Engineering and Tehnology, Kolkata, 700023, India

More information

Improved Vehicle Classification in Long Traffic Video by Cooperating Tracker and Classifier Modules

Improved Vehicle Classification in Long Traffic Video by Cooperating Tracker and Classifier Modules Improved Vehile Classifiation in Long Traffi Video by Cooperating Traker and Classifier Modules Brendan Morris and Mohan Trivedi University of California, San Diego San Diego, CA 92093 {b1morris, trivedi}@usd.edu

More information

To Do. Assignment Overview. Outline. Mesh Viewer (3.1) Mesh Connectivity (3.2) Advanced Computer Graphics (Spring 2013)

To Do. Assignment Overview. Outline. Mesh Viewer (3.1) Mesh Connectivity (3.2) Advanced Computer Graphics (Spring 2013) daned Computer Graphis (Spring 23) CS 283, Leture 5: Mesh Simplifiation Rai Ramamoorthi http://inst.ees.berkeley.edu/~s283/sp3 To Do ssignment, Due Feb 22 Start reading and working on it now. This leture

More information

Time delay estimation of reverberant meeting speech: on the use of multichannel linear prediction

Time delay estimation of reverberant meeting speech: on the use of multichannel linear prediction University of Wollongong Researh Online Faulty of Informatis - apers (Arhive) Faulty of Engineering and Information Sienes 7 Time delay estimation of reverberant meeting speeh: on the use of multihannel

More information

ALPHA-TRIMMED MEAN RADIAL BASIS FUNCTIONS AND THEIR APPLICATION IN OBJECT MODELING

ALPHA-TRIMMED MEAN RADIAL BASIS FUNCTIONS AND THEIR APPLICATION IN OBJECT MODELING ALPHA-TRIMMED MEAN RADIAL BASIS FUNCTIONS AND THEIR APPLICATION IN OBJECT MODELING Adrian G. Boq Ioannis Pitas Department University Thessaloniki of Informatis of Thessaloniki 54006, Greee ABSTRACT In

More information

FOREGROUND OBJECT EXTRACTION USING FUZZY C MEANS WITH BIT-PLANE SLICING AND OPTICAL FLOW

FOREGROUND OBJECT EXTRACTION USING FUZZY C MEANS WITH BIT-PLANE SLICING AND OPTICAL FLOW FOREGROUND OBJECT EXTRACTION USING FUZZY C EANS WITH BIT-PLANE SLICING AND OPTICAL FLOW SIVAGAI., REVATHI.T, JEGANATHAN.L 3 APSG, SCSE, VIT University, Chennai, India JRF, DST, Dehi, India. 3 Professor,

More information

A Partial Sorting Algorithm in Multi-Hop Wireless Sensor Networks

A Partial Sorting Algorithm in Multi-Hop Wireless Sensor Networks A Partial Sorting Algorithm in Multi-Hop Wireless Sensor Networks Abouberine Ould Cheikhna Department of Computer Siene University of Piardie Jules Verne 80039 Amiens Frane Ould.heikhna.abouberine @u-piardie.fr

More information

Evolutionary Feature Synthesis for Image Databases

Evolutionary Feature Synthesis for Image Databases Evolutionary Feature Synthesis for Image Databases Anlei Dong, Bir Bhanu, Yingqiang Lin Center for Researh in Intelligent Systems University of California, Riverside, California 92521, USA {adong, bhanu,

More information

Contents Contents...I List of Tables...VIII List of Figures...IX 1. Introduction Information Retrieval... 8

Contents Contents...I List of Tables...VIII List of Figures...IX 1. Introduction Information Retrieval... 8 Contents Contents...I List of Tables...VIII List of Figures...IX 1. Introdution... 1 1.1. Internet Information...2 1.2. Internet Information Retrieval...3 1.2.1. Doument Indexing...4 1.2.2. Doument Retrieval...4

More information

Analysis of input and output configurations for use in four-valued CCD programmable logic arrays

Analysis of input and output configurations for use in four-valued CCD programmable logic arrays nalysis of input and output onfigurations for use in four-valued D programmable logi arrays J.T. utler H.G. Kerkhoff ndexing terms: Logi, iruit theory and design, harge-oupled devies bstrat: s in binary,

More information

New Fuzzy Object Segmentation Algorithm for Video Sequences *

New Fuzzy Object Segmentation Algorithm for Video Sequences * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 521-537 (2008) New Fuzzy Obet Segmentation Algorithm for Video Sequenes * KUO-LIANG CHUNG, SHIH-WEI YU, HSUEH-JU YEH, YONG-HUAI HUANG AND TA-JEN YAO Department

More information

Prototype Based Classifier Design with Pruning

Prototype Based Classifier Design with Pruning Prototype Based Classifier Design with Pruning Jiang Li, Michael T. Manry, and Changhua Yu Department of Electrical Engineering, University of Texas at Arlington Arlington, Texas 7619. manry@uta.edu Abstract

More information

Acoustic Links. Maximizing Channel Utilization for Underwater

Acoustic Links. Maximizing Channel Utilization for Underwater Maximizing Channel Utilization for Underwater Aousti Links Albert F Hairris III Davide G. B. Meneghetti Adihele Zorzi Department of Information Engineering University of Padova, Italy Email: {harris,davide.meneghetti,zorzi}@dei.unipd.it

More information

An Alternative Approach to the Fuzzifier in Fuzzy Clustering to Obtain Better Clustering Results

An Alternative Approach to the Fuzzifier in Fuzzy Clustering to Obtain Better Clustering Results An Alternative Approah to the Fuzziier in Fuzzy Clustering to Obtain Better Clustering Results Frank Klawonn Department o Computer Siene University o Applied Sienes BS/WF Salzdahlumer Str. 46/48 D-38302

More information

with respect to the normal in each medium, respectively. The question is: How are θ

with respect to the normal in each medium, respectively. The question is: How are θ Prof. Raghuveer Parthasarathy University of Oregon Physis 35 Winter 8 3 R EFRACTION When light travels from one medium to another, it may hange diretion. This phenomenon familiar whenever we see the bent

More information

Exploiting Enriched Contextual Information for Mobile App Classification

Exploiting Enriched Contextual Information for Mobile App Classification Exploiting Enrihed Contextual Information for Mobile App Classifiation Hengshu Zhu 1 Huanhuan Cao 2 Enhong Chen 1 Hui Xiong 3 Jilei Tian 2 1 University of Siene and Tehnology of China 2 Nokia Researh Center

More information

One Against One or One Against All : Which One is Better for Handwriting Recognition with SVMs?

One Against One or One Against All : Which One is Better for Handwriting Recognition with SVMs? One Against One or One Against All : Whih One is Better for Handwriting Reognition with SVMs? Jonathan Milgram, Mohamed Cheriet, Robert Sabourin To ite this version: Jonathan Milgram, Mohamed Cheriet,

More information

Adapting K-Medians to Generate Normalized Cluster Centers

Adapting K-Medians to Generate Normalized Cluster Centers Adapting -Medians to Generate Normalized Cluster Centers Benamin J. Anderson, Deborah S. Gross, David R. Musiant Anna M. Ritz, Thomas G. Smith, Leah E. Steinberg Carleton College andersbe@gmail.om, {dgross,

More information

Unsupervised Stereoscopic Video Object Segmentation Based on Active Contours and Retrainable Neural Networks

Unsupervised Stereoscopic Video Object Segmentation Based on Active Contours and Retrainable Neural Networks Unsupervised Stereosopi Video Objet Segmentation Based on Ative Contours and Retrainable Neural Networks KLIMIS NTALIANIS, ANASTASIOS DOULAMIS, and NIKOLAOS DOULAMIS National Tehnial University of Athens

More information

System-Level Parallelism and Throughput Optimization in Designing Reconfigurable Computing Applications

System-Level Parallelism and Throughput Optimization in Designing Reconfigurable Computing Applications System-Level Parallelism and hroughput Optimization in Designing Reonfigurable Computing Appliations Esam El-Araby 1, Mohamed aher 1, Kris Gaj 2, arek El-Ghazawi 1, David Caliga 3, and Nikitas Alexandridis

More information

What are Cycle-Stealing Systems Good For? A Detailed Performance Model Case Study

What are Cycle-Stealing Systems Good For? A Detailed Performance Model Case Study What are Cyle-Stealing Systems Good For? A Detailed Performane Model Case Study Wayne Kelly and Jiro Sumitomo Queensland University of Tehnology, Australia {w.kelly, j.sumitomo}@qut.edu.au Abstrat The

More information

Plot-to-track correlation in A-SMGCS using the target images from a Surface Movement Radar

Plot-to-track correlation in A-SMGCS using the target images from a Surface Movement Radar Plot-to-trak orrelation in A-SMGCS using the target images from a Surfae Movement Radar G. Golino Radar & ehnology Division AMS, Italy ggolino@amsjv.it Abstrat he main topi of this paper is the formulation

More information

On the Generation of Multiplexer Circuits for Pass Transistor Logic

On the Generation of Multiplexer Circuits for Pass Transistor Logic Preprint from Proeedings of DATE 2, Paris, rane, Marh 2 On the Generation of Multiplexer Ciruits for Pass Transistor Logi Christoph Sholl Bernd Beker Institute of Computer Siene Albert Ludwigs University

More information

A Descriptive Framework for the Multidimensional Medical Data Mining and Representation

A Descriptive Framework for the Multidimensional Medical Data Mining and Representation Journal of Computer Siene 7 (4): 519-55, 011 ISSN 1549-3636 011 Siene Publiations A Desriptive Framework for the Multidimensional Medial Data Mining and Representation Veeramalai Sankaradass and Kannan

More information

Naïve Bayes Slides are adapted from Sebastian Thrun (Udacity ), Ke Chen Jonathan Huang and H. Witten s and E. Frank s Data Mining and Jeremy Wyatt,

Naïve Bayes Slides are adapted from Sebastian Thrun (Udacity ), Ke Chen Jonathan Huang and H. Witten s and E. Frank s Data Mining and Jeremy Wyatt, Naïve Bayes Slides are adapted from Sebastian Thrun (Udaity ), Ke Chen Jonathan Huang and H. Witten s and E. Frank s Data Mining and Jeremy Wyatt, Bakground There are three methods to establish a lassifier

More information

Detecting Moving Targets in Clutter in Airborne SAR via Keystoning and Multiple Phase Center Interferometry

Detecting Moving Targets in Clutter in Airborne SAR via Keystoning and Multiple Phase Center Interferometry Deteting Moving Targets in Clutter in Airborne SAR via Keystoning and Multiple Phase Center Interferometry D. M. Zasada, P. K. Sanyal The MITRE Corp., 6 Eletroni Parkway, Rome, NY 134 (dmzasada, psanyal)@mitre.org

More information

Semantic Concept Detection Using Weighted Discretization Multiple Correspondence Analysis for Disaster Information Management

Semantic Concept Detection Using Weighted Discretization Multiple Correspondence Analysis for Disaster Information Management Semanti Conept Detetion Using Weighted Disretization Multiple Correspondene Analysis for Disaster Information Management Samira Pouyanfar and Shu-Ching Chen Shool of Computing and Information Sienes Florida

More information

Partial Character Decoding for Improved Regular Expression Matching in FPGAs

Partial Character Decoding for Improved Regular Expression Matching in FPGAs Partial Charater Deoding for Improved Regular Expression Mathing in FPGAs Peter Sutton Shool of Information Tehnology and Eletrial Engineering The University of Queensland Brisbane, Queensland, 4072, Australia

More information

Segmentation of brain MR image using fuzzy local Gaussian mixture model with bias field correction

Segmentation of brain MR image using fuzzy local Gaussian mixture model with bias field correction IOSR Journal of VLSI and Signal Proessing (IOSR-JVSP) Volume 2, Issue 2 (Mar. Apr. 2013), PP 35-41 e-issn: 2319 4200, p-issn No. : 2319 4197 Segmentation of brain MR image using fuzzy loal Gaussian mixture

More information

The Happy Ending Problem

The Happy Ending Problem The Happy Ending Problem Neeldhara Misra STATUTORY WARNING This doument is a draft version 1 Introdution The Happy Ending problem first manifested itself on a typial wintery evening in 1933 These evenings

More information

On - Line Path Delay Fault Testing of Omega MINs M. Bellos 1, E. Kalligeros 1, D. Nikolos 1,2 & H. T. Vergos 1,2

On - Line Path Delay Fault Testing of Omega MINs M. Bellos 1, E. Kalligeros 1, D. Nikolos 1,2 & H. T. Vergos 1,2 On - Line Path Delay Fault Testing of Omega MINs M. Bellos, E. Kalligeros, D. Nikolos,2 & H. T. Vergos,2 Dept. of Computer Engineering and Informatis 2 Computer Tehnology Institute University of Patras,

More information

High-Precision Numerical Scheme for Vortical Flow

High-Precision Numerical Scheme for Vortical Flow Applied Mathematis, 013, 4, 17-5 http://dx.doi.org/10.436/am.013.410a1004 Published Online Otober 013 (http://www.sirp.org/journal/am) High-Preision Numerial Sheme for Vortial Flow Kei Ito 1*, Tomoaki

More information

Detecting Outliers in High-Dimensional Datasets with Mixed Attributes

Detecting Outliers in High-Dimensional Datasets with Mixed Attributes Deteting Outliers in High-Dimensional Datasets with Mixed Attributes A. Koufakou, M. Georgiopoulos, and G.C. Anagnostopoulos 2 Shool of EECS, University of Central Florida, Orlando, FL, USA 2 Dept. of

More information

FUZZY WATERSHED FOR IMAGE SEGMENTATION

FUZZY WATERSHED FOR IMAGE SEGMENTATION FUZZY WATERSHED FOR IMAGE SEGMENTATION Ramón Moreno, Manuel Graña Computational Intelligene Group, Universidad del País Vaso, Spain http://www.ehu.es/winto; {ramon.moreno,manuel.grana}@ehu.es Abstrat The

More information

Self-Adaptive Parent to Mean-Centric Recombination for Real-Parameter Optimization

Self-Adaptive Parent to Mean-Centric Recombination for Real-Parameter Optimization Self-Adaptive Parent to Mean-Centri Reombination for Real-Parameter Optimization Kalyanmoy Deb and Himanshu Jain Department of Mehanial Engineering Indian Institute of Tehnology Kanpur Kanpur, PIN 86 {deb,hjain}@iitk.a.in

More information

Particle Swarm Optimization for the Design of High Diffraction Efficient Holographic Grating

Particle Swarm Optimization for the Design of High Diffraction Efficient Holographic Grating Original Artile Partile Swarm Optimization for the Design of High Diffration Effiient Holographi Grating A.K. Tripathy 1, S.K. Das, M. Sundaray 3 and S.K. Tripathy* 4 1, Department of Computer Siene, Berhampur

More information

Improved Circuit-to-CNF Transformation for SAT-based ATPG

Improved Circuit-to-CNF Transformation for SAT-based ATPG Improved Ciruit-to-CNF Transformation for SAT-based ATPG Daniel Tille 1 René Krenz-Bååth 2 Juergen Shloeffel 2 Rolf Drehsler 1 1 Institute of Computer Siene, University of Bremen, 28359 Bremen, Germany

More information

3-D IMAGE MODELS AND COMPRESSION - SYNTHETIC HYBRID OR NATURAL FIT?

3-D IMAGE MODELS AND COMPRESSION - SYNTHETIC HYBRID OR NATURAL FIT? 3-D IMAGE MODELS AND COMPRESSION - SYNTHETIC HYBRID OR NATURAL FIT? Bernd Girod, Peter Eisert, Marus Magnor, Ekehard Steinbah, Thomas Wiegand Te {girod eommuniations Laboratory, University of Erlangen-Nuremberg

More information

Polyhedron Volume-Ratio-based Classification for Image Recognition

Polyhedron Volume-Ratio-based Classification for Image Recognition Polyhedron Volume-Ratio-based Classifiation for Image Reognition Qingxiang Feng, Jeng-Shyang Pan, Senior Member, IEEE, Jar-Ferr Yang, Fellow, IEEE and Yang-ing Chou Abstrat In this paper, a novel method,

More information

Department of Electrical and Computer Engineering University of Wisconsin Madison. Fall

Department of Electrical and Computer Engineering University of Wisconsin Madison. Fall Department of Eletrial and Computer Engineering University of Wisonsin Madison ECE 553: Testing and Testable Design of Digital Systems Fall 2014-2015 Assignment #2 Date Tuesday, September 25, 2014 Due

More information

Multi-Piece Mold Design Based on Linear Mixed-Integer Program Toward Guaranteed Optimality

Multi-Piece Mold Design Based on Linear Mixed-Integer Program Toward Guaranteed Optimality INTERNATIONAL CONFERENCE ON MANUFACTURING AUTOMATION (ICMA200) Multi-Piee Mold Design Based on Linear Mixed-Integer Program Toward Guaranteed Optimality Stephen Stoyan, Yong Chen* Epstein Department of

More information

Video Data and Sonar Data: Real World Data Fusion Example

Video Data and Sonar Data: Real World Data Fusion Example 14th International Conferene on Information Fusion Chiago, Illinois, USA, July 5-8, 2011 Video Data and Sonar Data: Real World Data Fusion Example David W. Krout Applied Physis Lab dkrout@apl.washington.edu

More information

DECODING OF ARRAY LDPC CODES USING ON-THE FLY COMPUTATION Kiran Gunnam, Weihuang Wang, Euncheol Kim, Gwan Choi, Mark Yeary *

DECODING OF ARRAY LDPC CODES USING ON-THE FLY COMPUTATION Kiran Gunnam, Weihuang Wang, Euncheol Kim, Gwan Choi, Mark Yeary * DECODING OF ARRAY LDPC CODES USING ON-THE FLY COMPUTATION Kiran Gunnam, Weihuang Wang, Eunheol Kim, Gwan Choi, Mark Yeary * Dept. of Eletrial Engineering, Texas A&M University, College Station, TX-77840

More information

Incremental Mining of Partial Periodic Patterns in Time-series Databases

Incremental Mining of Partial Periodic Patterns in Time-series Databases CERIAS Teh Report 2000-03 Inremental Mining of Partial Periodi Patterns in Time-series Dataases Mohamed G. Elfeky Center for Eduation and Researh in Information Assurane and Seurity Purdue University,

More information

Chemical, Biological and Radiological Hazard Assessment: A New Model of a Plume in a Complex Urban Environment

Chemical, Biological and Radiological Hazard Assessment: A New Model of a Plume in a Complex Urban Environment hemial, Biologial and Radiologial Haard Assessment: A New Model of a Plume in a omplex Urban Environment Skvortsov, A.T., P.D. Dawson, M.D. Roberts and R.M. Gailis HPP Division, Defene Siene and Tehnology

More information

CONSIDERING the high computational load of synthetic

CONSIDERING the high computational load of synthetic IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 5, NO. 4, OCTOBER 2008 573 Doppler Keystone Transform: An Approah Suitable for Parallel Implementation of SAR Moving Target Imaging Gang Li, Member, IEEE,

More information

Seismic Refraction Method: A Technique for Determining the Thickness of Stratified Substratum

Seismic Refraction Method: A Technique for Determining the Thickness of Stratified Substratum Amerian Journal of Applied Sienes 0 (8: 857-86, 0 ISSN: 546-99 0 O. Anomohanran, his open aess artile is distributed under a Creatie Commons Attribution (CC-BY.0 liense doi:0.844/ajassp.0.857.86 Published

More information

Colouring contact graphs of squares and rectilinear polygons de Berg, M.T.; Markovic, A.; Woeginger, G.

Colouring contact graphs of squares and rectilinear polygons de Berg, M.T.; Markovic, A.; Woeginger, G. Colouring ontat graphs of squares and retilinear polygons de Berg, M.T.; Markovi, A.; Woeginger, G. Published in: nd European Workshop on Computational Geometry (EuroCG 06), 0 Marh - April, Lugano, Switzerland

More information

arxiv: v1 [cs.gr] 10 Apr 2015

arxiv: v1 [cs.gr] 10 Apr 2015 REAL-TIME TOOL FOR AFFINE TRANSFORMATIONS OF TWO DIMENSIONAL IFS FRACTALS ELENA HADZIEVA AND MARIJA SHUMINOSKA arxiv:1504.02744v1 s.gr 10 Apr 2015 Abstrat. This work introdues a novel tool for interative,

More information

Detection of RF interference to GPS using day-to-day C/No differences

Detection of RF interference to GPS using day-to-day C/No differences 1 International Symposium on GPS/GSS Otober 6-8, 1. Detetion of RF interferene to GPS using day-to-day /o differenes Ryan J. R. Thompson 1#, Jinghui Wu #, Asghar Tabatabaei Balaei 3^, and Andrew G. Dempster

More information